Authors: Timothy D McNamara, Mba-tihssommah Mosore, Alexander Urlaub, Marcus A Lashley, Nathan D Burkett-Cadena, Lawrence E Reeves, Estelle M Martin
Categories: Vector/Pathogen/Host Interaction, Transmission, Gopherus polyphemus, Mosquito, arbovirus, overwintering, AcademicSubjects/SCI01382, AcademicSubjects/MED00860
Source: Journal of Medical Entomology
Doi: 10.1093/jme/tjad174
Mosquito-borne diseases represent a significant threat to human and animal health in the United States. Several viruses, including West Nile, Saint Louis encephalitis, and Eastern equine encephalitis are endemic. In humans, the disease is typically detected during the summer months, but not during the winter months. The ability of these viruses to reemerge year after year is still not fully understood, but typically involves persistence in a reservoir host or vector during periods of low transmission. Mosquito species are known to overwinter at different life stages (adults, larvae, or eggs) in manufactured or natural sites. Gopher tortoise burrows are known to serve as refuge for many vertebrate and invertebrate species in pine savannas. In this study, we surveyed the interior of gopher tortoise burrows for overwintering mosquitoes. We identified 4 species (Anopheles crucians s.l., Culex erraticus, Mansonia dyari, and Uranotaenia sapphirina). Cx. erraticus was the most abundant, and its presence and abundance increased in winter months, implying that this species utilized gopher tortoise burrows for overwintering. Bloodfed Cx. erraticus and An. crucians s.l. females were detected. While An. crucians s.l. fed exclusively on the white-tailed deer, Cx. erraticus had a more diverse host range but fed primarily on the gopher tortoise. Tortoises and other long-lived reptiles like the American alligator have been shown to sustain high viremia following West Nile virus (WNV) and Eastern equine encephalitis virus (EEEV) infection and therefore could play a role in the maintenance of these viruses. In addition, Cx. erraticus is naturally infected with WNV and is a known bridge vector for EEEV. As such, these overwintering sites may play a role in perpetuating over-winter arboviral activity in Florida.
Keywords: Mosquito, overwintering, arbovirus, Gopherus polyphemus
West Nile virus (WNV), Eastern equine encephalitis virus (EEEV), and Saint Louis encephalitis virus (SLEV) are endemic to the United States and represent a significant threat to human health (Soto 2022). While these viruses are enzootic, circulating among wild birds and ornithophilic mosquitoes, epidemics in humans or domestic animals are tied to feeding on nonpreferred mammalian hosts during periods of high arboviral transmission or shifts in mosquito feeding due to the migration of preferred host populations (Turell et al. 2005, Weaver 2005, Hamer et al. 2009). Since its introduction into the United States in 1999, WNV has caused more than 28,614 neuroinvasive disease cases, and since 2011, EEEV and SLEV are responsible for 186 and 219 neuroinvasive disease cases, respectively (CDC 2023a,b,c).
Transmission of these arboviruses is cyclical, with the emergence in reservoir and amplifying hosts in the spring, in human and domestic animals in the summer through fall, and lower transmission or disappearance in the winter when temperatures are low and adult mosquito populations decline (Rückert and Ebel 2018). The ability of these viruses to reemerge each year implies that they can persist through unfavorable periods for transmission. Several mechanisms have been identified, including persistence within a host that migrates to areas where the enzootic cycle can be maintained, transmission to a host capable of maintaining viremia over the winter, or persistence within the mosquito population as it overwinters (Watts et al. 1987, Beaty and Calisher 1991, Reisen and Wheeler 2019). In cases where viruses persist within the mosquitoes, maintenance occurs as a result of dormancy within infected individuals, either through diapause or quiescence. The physiological state of diapause is marked by a reduction or cessation in many biological processes, allowing the organism to persist during periods unsuitable for their typical life history. In mosquitoes, diapause is initiated by shortening daylength. The specific critical photoperiod for diapause initiation varies by species and latitude of the population; however, many members of the genus Culex enter diapause once daylength decreases below ~12 h. Similarly, the factors leading to the cessation of diapause vary but are often the result of environmental factors such as photoperiod and temperature. In addition to variation at the point in which diapause is initiated and exited, mosquitoes often diapause at different life stages, including egg (e.g., Aedes and Psorophora), larva (e.g., Culiseta and Coquillettidia), and adult (e.g., Culex and Anopheles) (Eldridge 1968, Bradshaw 1976, Denlinger and Armbruster 2014, Diniz et al. 2017). Many of the Culex vectors of WNV are known to overwinter as adults, with infected overwintering adults potentially playing a role in early season emergence through vertical transmission from infected females to their offspring (Eldridge 1987, Nasci et al. 2001, Dohm et al. 2002, Goddard et al. 2003, Rochlin et al. 2019). Adult mosquitoes are known to utilize the different habitats for diapause, including manufactured and natural sites. These sites are often cryptic, located in protected, dry, and dark microhabitats such as tree holes and culverts (Burkett-Cadena et al. 2008a, 2011a). This makes surveilling overwintering populations difficult, thus limiting our ability to assess the role overwintering plays in the early season establishment and the transmission of mosquito-borne viruses.
One potential method to surveil overwintering mosquitoes consistently is to seek out potential hosts that construct protected microhabitats of their own. For example, the gopher tortoise (Gopherus polyphemus) is an ecosystem engineer that constructs deep burrows in well-drained, sandy soils of the southeastern United States (Hansen 1963). The gopher tortoise is listed as a threatened reptile species in the United States due to habitat loss and is the target of conservation efforts, but populations have been rebounding in recent years and are stable throughout much of its eastern range (USFWS 2022). It is the most active above-ground from April to October and is associated with longleaf pine (Pinus palustris) savanna and sandhill ecosystems in Florida, Georgia, Alabama, Mississippi, Louisiana, and South Carolina (Auffenberg 1982, Catano and Stout 2015). The gopher tortoise is a keystone species that affects animal and plant communities as seed dispersers and by providing refuge sites for more than 300 species of invertebrates and 60 species of vertebrates in their burrows (Jackson 1989, Kent et al. 1997, McCoy et al. 2006, Dziadzio and Smith 2016). The burrows themselves are insulated from temperature changes and precipitation, creating refugia for small mammals and reptiles, which are not only potential hosts for mosquitoes but also overwintering sites for cold-sensitive species such as the eastern indigo snake and eastern diamondback rattlesnake (Bauder et al. 2017). Many long-lived reptiles sustain high viremia following WNV and EEEV infection and thus could play a role in the maintenance of these viruses overwinter (Bowen 1977, White et al. 2011, Byas et al. 2022).
As a result of both the presence of hosts and a stable microsite, these burrows may be beneficial to mosquitoes for overwintering. However, to the best of our knowledge, no prior published study has investigated the presence of mosquitoes in G. polyphemus burrows or their role in mosquito ecology. Nevertheless, because of their ability to create a high density of burrows in natural habitats across a broad geographic region where mosquito-vectored diseases are most prevalent in the United States, burrows may play an unrecognized role in harboring diseases overwinter.
To determine the role G. polyphemus burrows may play in mosquito overwintering, we surveyed 11 burrow sites at the Ordway−Swisher Biological Station in north-central Florida over 10 months using a modified Prokopack aspirator. Additionally, we surveyed burrows in abandoned pasture and sandhill ecosystems to determine whether mosquito usage of burrows was tied to specific habitats. Finally, we assessed the feeding status and conducted blood meal analyses of collected mosquitoes to assess the temporal variation of feeding and determine what groups of organisms they use as hosts.
Surveillance of G. polyphemus burrows occurred at the University of Florida’s Ordway−Swisher Biological Station (Lat: 29.691489, Long: −82.009476), a ~38.4 km^2^ research facility focused on the long-term research and conservation of ecosystems. While much of the station is made up of the sandhill ecosystem, a habitat consisting of well-drained sandy hills along with stands of longleaf pine (P. palustris), portions of the research station are degraded due to its use as pastureland prior to its acquisition by the University of Florida. Areas belonging to unmodified sand hill and abandoned pasture ecotypes were surveyed for the presence of G. polyphemus burrows to determine the sampling locations for mosquito surveillance (Fig. 1). From the identified candidate burrows, pairs of burrows within 30 m of each other and within the same ecotype were grouped with pairs in the other ecotype that were spaced at least 300 m away to form a single sampling site. Forty-four burrows were identified within the research area, allowing for 11 sites for mosquito sampling (Fig. 2).
Fig. 1. Abandoned Pasture (left) and Sandhill (right) habitats at the Ordway-Swisher Biological Station, University of Florida, IFAS, USA. Pair of images showing the differences between abandonned pasture and sanhilll habitat at the Ordway-Swisher Biological Station at the University of Florida. Abandonoed pasture habitat contain sandy soil, grasses and is largely devoided of trees while the Sandhill habitat is a mix of grasses and longleaf pines.
Fig. 2. Locations of G. polyphemus burrows used in mosquito sampling within the Ordway−Swisher Biological Station, University of Florida, IFAS, USA. Map of Florida showing the Ordway -Swisher biological station in the North Central region of the State along with the location of burrows used for mosquito surveillance.
From September 2021 to June 2022, mosquitoes were sampled once a month from G. polyphemus burrows using a Prokopack Aspirator (John W. Hock Company, Florida, United States) modified with a telescoping sleeve to access 2 m into the burrow interior (Fig. 3). Burrows were aspirated for 60 s, during which the aspirator was inserted from the burrow apron into the burrow tunnel as far as possible, collecting from the walls and ceiling of the burrow. Collected mosquitoes were stored in an insulated container with ) and transported to the University of Florida (Gainesville, FL; ~30 min from Ordway−Swisher Biological Station) for identification. Upon arrival, samples were stored at −80°C until further processing.1 kg of dry ice (CO2
Fig. 3. Adapted Prokopack fitted with a 2 feet plastic sleeve (top). The Prokopack aspirator was inserted at the entrance of the burrows and aspiration performed for 1 min to collect mosquitoes present inside gopher tortoise burrows (bottom). Pair of images showing the modified prokopack aspirator with a telescoping tube on the ground and in use by researcher to asprate the interior of a gopher tortoise burrow.
The collected mosquitoes were sorted by date, site, species, sex, and feeding state. Species identification was performed using a dichotomous key of mosquitoes from North America (Darsie and Ward 2005). Mosquitoes identified as bloodfed were stored separately in individual 1.5 ml vials at −80°C, and used in blood meal analysis. Mosquitoes identified as Anopheles crucians s.l. were identified to species level with DNA barcoding (Reeves et al. 2021).
Photoperiod measurements were collected using the National Oceanic and Atmospheric Administration (NOAA) Solar Calculator (NOAA 2005). Sunrise and sunset times for the Ordway−Swisher Biological Station during the 7 days immediately preceding the dates of mosquito collection were used to calculate an average photoperiod for each collection week. Average photoperiods were then assigned a binary designation based on whether they were less than 12 h in length (daylength) (Eldridge 1968).
Blood from bloodfed females was transferred onto a Flinders Technology Associates Classic Card (Sigma-Aldrich Corp., St. Louis, Missouri) using a sterile pipette tip and scored (1−3) based on the extent of digestion (Reeves et al. 2016, 2018a). The Hot Sodium Hydroxide and Tris (HotSHOT) methods were used to extract host DNA (Reeves and Burkett-Cadena 2023). Briefly, 30 µl of a lysis solution (25 mM sodium hydroxide and 0.2 mM ethylenediaminetetraacetic acid) were added to two 1 mm diameter punches of blood meal and incubated for 60 min at 95°C followed by 5 min at 4°C. Then, 30 µl of neutralizing solution (40 mM Tris-HCl) was added, and the DNA was stored at −20°C until further processing. Each sample was amplified using a hierarchical polymerase chain reaction (PCR) approach using 3 primer pairs targeting the cytochrome c oxidase subunit I (CO1) gene. Each sample was first amplified using the VertCO1-7194F/Mod_RepCO1_R primer pair, followed by the Mod-RepCO1_F/Vert CO1_7216R and the VertCO1_10096_F/Mod_RepCO1_R primer pairs in cases of failed amplicon production. Each PCR was performed using 2 µl of DNA extract, 1 µl of forward primer (10 µM), 1 µl of the reverse primer (10 µM), and 10 µl of the 2.0× Apex taq RED Master mix (Genesee Scientific Corp., San Diego, California), and 6 µl of ultrapure water. Amplification began with 1 cycle at 95°C for 3 min, followed by 40 cycles of 40 s at 95°C, 30 s at 45°C, 3 min at 72°C, and a final extension step for 7 min at 72°C. PCR products were visualized on a 1.5% agarose gel and amplicons sent to Eurofins Genomics (Louisville, Kentucky) for PCR clean-up and Sanger sequencing. Sequences were identified with the Barcode of Life Database.
To assess the usage of G. polyphemus burrows as mosquito diapause sites and the impact that habitat has on that usage, we conducted analyses of mosquito presence, abundance, and feeding state using generalized linear mixed modeling in SPSS. Models were generated by using the number of female Culex erraticus collected per burrow (abundance), the presence of female C. erraticus within burrows (presence), and the percentage of bloodfed C. erraticus females within burrows (feeding state) as response variables and ecotype, daylength, and the interaction between ecotype and daylength as explanatory variables. To account for site-level and sampling event variability, models included the site of collection and month of collection as random variables. Finally, to account for distributions within collected data, models were run using a log-linear distribution (for C. erraticus abundance), a binomial distribution (for C. erraticus presence), or a Gaussian distribution (for percent bloodfed C. erraticus). Based on the results of model analysis, significant factors containing only two variables were analyzed using model analysis results, while significant factors containing more than two variables (the interaction between ecotype and daylength) were conducted using pairwise analysis with a Bonferroni correction.
Over the 10 months of surveillance, 2,768 mosquitoes were collected from G. polyphemus burrows. The majority (97.3%) of collected mosquitoes were female. Mosquitoes belonged to 4 Cx. erraticus (98.0%), An. crucians s.l. (1.8%), Mansonia dyari (<0.1%), and Uranotaenia sapphirina (<0.1%). On average, 6.2 ± 0.7 (mean ± standard error [SE]) Cx. erraticus (6.0 ± 0.7 females and 0.2 ± 0.1 males), 0.1 ± 0.1 An. crucians s.l. (<0.1 ± 0.1 females and < 0.1 ± 0.1 males), <0.1 ± 0.1 Ma. dyari (all females), and < 0.1 ± 0.1 Ur. sapphirina (all females) were collected per burrow, with rates of collection varying by month (Table 1). Based on the low collection rate (<1 mosquito per burrow per month) for An. crucians s.l., Ma. dyari, and Ur. sapphirina, only Cx. erraticus collection was used in the further statistical abundance analysis.
During the study period, sampling occurred a total of 440 times (44 burrows × 10 collection times) across all G. polyphemus burrows. During this sampling, Cx. erraticus females were collected in 37.0% of all sampled burrows. However, the rate of detection of female Cx. erraticus within burrows varied from 0.0% to 68.2% depending upon the month of collection. Model analysis showed no difference (F1, 436 = 2.373, p = 0.124) in the prevalence of Cx. erraticus females in G. polyphemus burrows according to ecotype (41.4% in abandoned pasture burrows and 32.3% in sandhill burrows) but daylength was a significant factor (F1, 436 = 8.962, p = 0.003). 58.2% of the burrows were positive for Cx. erraticus when the average weekly daylength was <12 h, versus only 15.5% in months when the average daylength was >12 h. The interaction between habitat and daylength was not significant (F1, 436 = 2.373, p = 0.124) (Fig. 4).
Fig. 4. Female Cx. erraticus presence in G. polyphemus burrows along with average daylength.* represents a significant difference in the rate of presence within burrows. Figure showing the presence of Culex erraticus within gopher tortoise burrows along with the photoperiod by month. The figure shows higher prevalence of Culex erraticus in burrows when photoperiod was less than 12h.
A total of 2,659 female Cx. erraticus were collected from G. polyphemus burrows, with an average of 6.0 ± 0.7 collected across all burrows. However, average rates of Cx. erraticus collection from burrows ranged from 0.0 ± 0.0 to 17.3 ± 3.7 depending on the month of collection. Similarly, average photoperiod varied from 10.25 to 14.05 h depending on the month of collection. Model analysis of Cx. erraticus female collection in burrows showed no difference (F1, 435 = 0.088, p = 0.293) between abandoned pasture (6.7 ± 0.5) and sandhill (5.4 ± 0.4) habitats. Average daylength was significant (F1, 435 = 15.412, p < 0.001), with 11.5 ± 1.3 (Mean ± SE) female Cx. erraticus collected from burrows when daylength was shorter than 12 h and 0.6 ± 0.1 collected when daylength was longer than 12 h. The interaction between daylength and burrow ecotype was significant (F1, 435 = 18.713, p < 0.001). However, Bonferroni-adjusted analysis of the impact of ecotype given daylength showed no significance (F1, 435 = 1.605, p = 0.206), with an average of 12.9 ± 1.9 and 10.1 ± 1.8 female Cx. erraticus collected from abandoned pasture and sandhill burrows (respectively) when daylength was <12 h, and 0.4 ± 0.1 and 0.7 ± 0.3 collected when daylength was >12 h (Fig. 5).
Fig. 5. Female Cx. erraticus abundance in G. polyphemus burrows based on month of collection, ecotype (Abandoned Pasture and Sandhill), and average daylength. * represents a significant difference in the abundance of Cx. erraticus within burrows. Figure showing the abundance of Culex erraticus within gopher tortoise burrows along with the photoperiod by month and habitat. The figure shows higher abundance of Culex erraticus in burrows when photoperiod was less than 12h. No difference in abundance was observed according to habitat.
A total of 40 bloodfed C. erraticus were collected over the course of sampling. Among collected Cx. erraticus females, bloodfed individuals made up 9.3 ± 2.0% of all samples.
However, the frequency of bloodfed Cx. erraticus collected from individual burrows varied between 0.0% and 100%, and between 0.0% and 41.1% across Cx. erraticus collected within a given month. Model analysis of the percentage of bloodfed Cx. erraticus collected from burrows showed no difference (F1, 159 = 0.032, p = 0.858) between abandoned pasture (8.7 ± 2.6%) and sandhill (10.0 ± 3.2%) burrows. Average daylength was significant (F1, 159 = 92.691, p < 0.001), with bloodfed females making up 1.3 ± 0.9% of Cx. erraticus females collected from burrows when daylength was shorter than 12 h, and 39.7 ± 7.0% when daylength was >12 h. The interaction between ecotype and daylength was not significant (F1, 159 = 0.466, p = 0.496), with bloodfed females making up 2.1 ± 1.5% and 0.1 ± 0.1% of female Cx. erraticus collected from abandoned pasture and sandhill burrows (respectively) during periods when average daylength was <12 h, and 37.7 ± 9.6% and 41.7 ± 10.5% of abandoned pasture and sandhill burrows (respectively) when average daylength was >12 h (Fig. 6).
Fig. 6. Average percentage of bloodfed Cx. erraticus in G. polyphemus burrows and total female Cx. erraticus collection by month of collection, grouped by daylength.* represents a significant difference in the average percentage of bloodfed Cx. erraticus. Figure showing Culex erraticus abundance withing burrows along with percentage of bloodfed individuals by month of collection. The figure shows higher rates of bloodfed mosquitoes during month when abundance was low and photoperiod was more than 12h.
In total, 55 bloodfed mosquitoes were collected (40 Cx. erraticus and 15 An. crucians s.l.). Fourteen of our bloodfed mosquitoes had fresh blood meal and were scored as 1, 12 had an older blood meal and were scored as 2 while 29 had degraded blood meals and were scored as 3. While we observed a high amplification success for our mosquitoes scored as 1 (78.57%) and 2 (75%), the amplification success significantly dropped to 10.34 % for mosquitoes scored as 3 which represented 52.72% of our samples. Out of the 40 Cx. erraticus tested in our blood meal analysis, 20 (50%) were successfully amplified. Sixteen of the amplified blood meals were identified as G. polyphemus (80.0%), 1 as Alligator mississipensis (American alligator) (5%), 1 as Odocoileus virginianus (white-tailed deer) (5%), 1 as Spizella passerina (chipping sparrow) (5%), and 1 as Setophaga americana (northern parula) (5%). Among successfully amplified blood meals taken from Cx. erraticus, 12 of the G. polyphemus samples, the 1 A. mississipensis sample, and the 1 O. virginianus sample were from months when daylength was >12 h. The remaining 4 G. polyphemus samples, the 1 S. passerina samples, and the 1 S. americana sample were obtained between October 2021 and March 2022, when daylength was shorter than 12 h (Table 2). Out of the 15 An. crucians s.l. tested for mosquito species identification, 1 was identified as An. crucians A and 14 as An. crucians D. While the blood meal from the An. crucians A sample could not be amplified, 3 blood meals (21.4%) were successfully amplified from the An. crucians D sub-set and identified as belonging to O. virginianus (Fig. 7). These 3 blood meals were collected in September.
Fig. 7. Vertebrate species identified from the mosquito blood meals for Cx. erraticus and An. crucians D in Alachua, Florida, USA Pie charts of the bloodmeal amplification from Culex erraticus and Anopheles crucians D. 50% of Culex erraticus samples did not amplify, 40% belonged to Gopherus polyphemus. The remaining hosts were identified as Spizella passerina, Odocoileus virginanus, Alligator mississipiensis, and Setophaga americana. 78.6% of Anopheles crucians D samples did not amplify, with the remaining belonging to Odocoileus virginanus.
The aim of this study was to assess whether mosquito species are present within and utilize G. polyphemus burrows either by feeding on the organisms that are associated with them or by using the burrows as refugia. We identified 4 species within burrows (Cx. erraticus, An. crucians s.l., Ma. dyari, and Ur. sapphirina). However, the abundance of collected mosquitoes was not even, implying that some species may be more directly associated with G. polyphemus burrows than others.
Among the 4 mosquito species detected, Cx. erraticus was the most prevalent (98% of all mosquitoes collected) species collected and burrow occupancy was observed to have a strong relationship with daylength. The highest prevalence and abundance were observed from October through February, when the average daylength was below 12 h and above-ground conditions are likely to be suboptimal for development and breeding. Many Culex mosquitoes overwinter as adults including Cx. erraticus (Eldridge 1987). Additionally, Cx. erraticus is known to overwinter in natural sites (tree hollows) in addition to manufactured sites (Burkett-Cadena et al. 2008a, 2011a). Daylength and other environmental cues including temperature are known to play a role in initiating and ending mosquito diapause. For example, daylength is known to trigger diapause in Culex pipiens (Gill et al. 2017). In Alabama, the emergence of Cx. erraticus from overwintering sites was observed in March and was associated with temperatures ranging from 18 to 22 °C (Burkett-Cadena et al. 2011a). While our study did not assess temperature, we observed that Cx. erraticus exited the burrows in March. It is possible that temperature may play a role in when Cx. erraticus leave burrows, as they may not receive sufficient photoperiod cues while underground. Further research will be necessary to determine how the temperature impacts Cx. erraticus overwintering in G. polyphemus burrows. Regardless, our findings likely represent a true observation of overwintering behavior in Cx. erraticus with shorter daylength triggering the increased occupancy of burrows.
No relationship between the presence or abundance of Cx. erraticus and the percentage of bloodfed Cx. erraticus was observed based on the habitat type. This is interesting, as it implies that one of the several potential factors is present. It is possible that the abandoned pasture and sandhill ecotypes had no difference in the Cx. erraticus abundance, despite differences in the composition and structure of the vegetation. Alternatively, factors present within the ecotypes may result in differences in aboveground Cx. erraticus abundance, or in Cx. erraticus occupancy within existing burrows which may have impacted the overall Cx. erraticus presence within burrows. It should be noted that we observed a higher abundance in G. polyphemus burrow density in abandoned pastures in comparison to sand hills in our study area. As such, it is possible that an increased abundance of burrows provided more opportunities for Cx. erraticus to disperse within abandoned pasture systems, resulting in an overall reduction in their presence within burrows relative to their abundance within the environment above. However, without additional research examining above- and below-ground abundance, it is impossible to whether habitat has any meaningful impact on Cx. erraticus presence and usage depending on ecotype.
Blood meal analysis revealed a strong inclination of Cx. erraticus toward G. polyphemus (80% of all amplified blood meals). Other hosts identified were A. mississipensis (5%), S. passerina (5%), S. americana (5%), and O. virginianus (5%). We noted a reduction, though not complete disappearance, in bloodfed Cx. erraticus during the period when daylength was <12 h. This implies that Cx. erraticus was using burrows mainly as resting sites during the overwintering period. However, it should be noted that 6 of our samples were collected during the time period, including the samples from S. passerina and S. americana which were obtained in November and December.
Our results support prior studies showing the broad range of hosts utilized by Cx. erraticus, which include mammals, amphibians, birds, and reptiles (Cupp et al. 2004, Rodrigues and Maruniak 2006, Watts et al. 2009). Above-ground studies taking place during the typical mosquito season have demonstrated the temporal pattern of Cx. erraticus with more blood meals taken from birds in spring and early summer and a shift to mammal feeding during the late summer and fall (Burkett-Cadena et al. 2008b). This temporality in feeding preference was further linked to host reproductive phenology for birds and mammals (Burkett-Cadena et al. 2011b). Previously, the winter-feeding preference of above-ground populations of Cx. erraticus was investigated in Hillsborough County, Florida showing preferential feeding on wading birds including the black-crowned night heron (Nycticorax nycticorax), the great blue heron (Ardea herodias), the great egret (Ardea alba), the wood stork (Mycteria americana) and the yellow-crowned night heron (Nyctanassa violacea) followed by mammals including humans and white-tailed deer. The lack of feeding on reptiles and amphibians was hypothesized to be related to brumation (Bingham et al. 2014). Our data provide further evidence to support this hypothesis given the high rate of G. polyphemus detection among Cx. erraticus blood meals during the spring and summer months when G. polyphemus are active, and the high prevalence and abundance of Cx. erraticus in burrows during the G. polyphemus brumation period in Florida. Additionally, while we did not observe any feeding on wading birds, both of our bird samples were observed during winter months. Thus, while our results vary, they do match the trends in feeding shifts observed previously.
Because of the high frequency of individuals found naturally infected in nature, the overlap in geographic distribution, and its vector competence, Cx. erraticus is a known bridge vector of EEEV. Cx. erraticus is also a suspected vector of WNV in the southern United States, but its vector competence for WNV remains untested to date due to the difficulties in rearing this species in laboratory settings. Our blood meal results may strengthen the hypotheses that Cx. erraticus plays a role in overwinter transmission of arboviruses such as WNV and EEEV, as several of our detected blood meals came from species that are known to be, or are potentially, competent reservoirs of arboviruses and occurred in the winter months. Notably, members of the Order Passeriformes, such as S. passerina, are known to present with sufficient EEEV and WNV viremia to infect susceptible mosquitoes (Komar et al. 1999, 2003, van der Meulen et al. 2005). Spizella passerina is found throughout North America and is known to exist in both migrant and resident populations dependent on their geographic location. Migrant populations are typically observed in the northern regions and migrate toward southern states, including Florida, from late Fall to early Spring (Middleton 2020). This can have implications for arboviral transmission patterns as S. passerina has been identified as a preferred host of Culiseta melanura, the primary vector of EEEV in the northeast. If S. passerina is positive for EEEV or WNV before migrating north, it is possible that this species plays a role in the reemergence of these viruses from southern to northern regions on a yearly basis (Molaei et al. 2016, Armstrong and Andreadis 2022). S. americana is known to migrate from southern Florida and Central and South America northwards (Moldenhauer and Regelski 2020). S. americana has been found naturally infected with WNV (Dusek et al. 2009, CDC 2016). Additionally, WNV has been detected in Alligator mississippiensis and has been shown to reach viremia sufficient for the infection of Culex quinquefasciatus, Cx. pipiens, and Culex tarsalis (Byas et al. 2022, Jacobson et al. 2005). Finally, the majority of the Cx. erraticus blood meals were from G. polyphemus, highlighting their possible frequent exposure to WNV and EEEV. While the competence status of G. polyphemus as a reservoir host is unknown, other reptiles (including members of the genus Gopherus) are potential reservoirs of EEEV. Garter snakes (Thamnophis sirtalis), green anoles (Anolis carolinensis), and the Texas tortoise (Gopherus berlandieri) have been shown to reach sufficient viremia for the transmission of EEEV, with the viremia persisting in both G. berlandieri and T. sirtalis for extended periods (Bowen 1977, White et al. 2011). As such, it is possible that Cx. erraticus may be involved in overwintering arboviruses such as WNV and EEEV by contracting it from hosts prior to overwintering. This could either take the form of persistence within overwintering adults that continue transmission when they become active in the spring, or by means of vertical transmission to offspring produced after exiting diapause. It is also possible that the species whose burrows are used for Cx. erraticus overwintering, such as G. polyphemus, may help the overwinter to these arboviruses by allowing them to persist until spring. However, little is known regarding the process of diapause in Cx. erraticus due to the difficulties in rearing them or previous challenges in collecting them in abundance from the environment. This has also made it difficult to determine how frequently or effectively Cx. erraticus vertically transmit pathogens to offspring. Finally, our present study did not surveil rates of arboviruses in collected Cx. erraticus or in G. polyphemus. Additional research will be necessary to determine the dynamics of arbovirus persistence within the overwintering Cx. erraticus populations.
Aside from Cx. erraticus, the occurrence of other mosquito species in burrows was rare and incidental for 2 of the species, rather than the result of an existing relationship with G. polyphemus burrows. For example, a single unfed female Ur. sapphirina was collected during sampling. This low rate of collection implies that there is not an association between Ur. sapphirina and G. polyphemus burrows, and that its presence was the result of host-seeking for annelids or the incidental use of the burrow as a resting site (Reeves et al. 2018b). Similarly, 2 Ma. dyari specimens were collected during the whole sampling period. Ma. dyari is known to develop attached to aquatic plants, including water lettuce, water hyacinth, and cattail, all known to be present in the pond adjacent to the site where the specimens were collected (Slaff and Haefner 1985). The collection of these 2 species during sampling was incidental, rather than the result of an existing relationship with G. polyphemus burrows.
More An. crucians s.l. were collected than Ma. dyari and Ur. sapphirina (51 in total); however, the small sample size made analysis of their association with burrows dubious. Despite this, several observations can be made about the individuals collected. First, 54.8% of the females were bloodfed, and all successfully identified blood meals taken from An. crucians s.l. females originated from O. virginianus, implying that these mosquitoes utilize the burrows as resting sites following a blood meal rather than feeding inside the burrows. Our blood meal analysis confirms the finding of others showing a strong preference of An. crucians s.l. toward large mammals (Cohen et al. 2009, Watts et al. 2009, Escobar et al. 2020). Second, proportionally, more males (39.2%) were collected for An. crucians s.l. in comparison to other species (2.0% for Cx. erraticus, and 0.1% for both Ur. sapphirina and Ma. dyari). This could imply the proximity of a breeding site and that G. polyphemus burrows may serve as a site for mating, rather than overwintering or feeding site. Additionally, both male and female An. crucians s.l. were seen gathering around the entrance of G. polyphemus burrows but would disperse when the aspirator was inserted. Their association with G. polyphemus burrows may be more closely tied to the burrow entrance than the interior. As our experimental design was meant to assess the usage of the interior of G. polyphemus burrows, we likely failed to capture the true abundance and temporal variation of An. crucians s.l. near the entrance of burrows. Further research will be necessary to determine the true relationship between An. crucians s.l. have to G. polyphemus burrows.
Regardless of the need for additional research, our study is unique and provides the first example of mosquitoes using G. polyphemus burrows as refuge, adding to the 302 species of invertebrates previously reported in these burrows (Jackson 1989). The strong association of Cx. erraticus with G. polyphemus burrows, coupled with its feeding preference, shows that this species uses burrows for overwintering and host-seeking. The role of the gopher tortoise as an amplifying or diluting host for arboviruses including WNV and EEEV still needs to be investigated.
We thank Remy Powell, Danielle Cabezas, Kendall Long, Nicole Vargas, and Amal Al-Rajhi for assistance with field collections. We also credit Emily Ferrall from the wildlife Resources Division at the Georgia Department of Natural Resources for the gopher tortoise photo used to create our graphical abstract.
Timothy D McNamara, Department of Entomology and Nematology, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, FL, USA.
Mba-tihssommah Mosore, Department of Entomology and Nematology, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, FL, USA.
Alexander Urlaub, Florida Medical Entomology Laboratory, Department of Entomology and Nematology, Institute of Food and Agricultural Sciences, University of Florida, Vero Beach, FL, USA.
Marcus A Lashley, Department of Wildlife Ecology and Conservation, University of Florida, Gainesville, FL, USA.
Nathan D Burkett-Cadena, Florida Medical Entomology Laboratory, Department of Entomology and Nematology, Institute of Food and Agricultural Sciences, University of Florida, Vero Beach, FL, USA.
Lawrence E Reeves, Florida Medical Entomology Laboratory, Department of Entomology and Nematology, Institute of Food and Agricultural Sciences, University of Florida, Vero Beach, FL, USA.
Estelle M Martin, Department of Entomology and Nematology, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, FL, USA.
This study was supported in part by the intramural research program of the U.S. Department of Agriculture, National Institute of Food and Agrculture, Hatch Regular #1024853. The findings and conclusions in this publication have not been formally disseminated by the U.S. Department of Agriculture and should not be construed to represent any agency determination or policy.
Timothy McNamara (Formal analysis [Lead], Visualization [Equal], Writing—original draft [Lead]), Mba-tihssommah Mosore (Data curation [Lead], Writing—review & editing [Supporting]), Alexander Urlaub (Data curation [Supporting], Writing—review & editing [Supporting]), Marcus Lashley (Methodology [Supporting], Resources [Supporting], Writing—review & editing [Supporting]), Nathan Burkett-Cadena (Methodology [Supporting], Writing—review & editing [Supporting]), Lawrence Reeves (Data curation [Supporting], Methodology [Supporting], Writing—review & editing [Supporting]), and Estelle Martin (Conceptualization [Lead], Data curation [Supporting], Formal analysis [Supporting], Funding acquisition [Lead], Investigation [Lead], Methodology [Lead], Project administration [Lead], Supervision [Lead], Visualization [Equal], Writing—original draft [Supporting], Writing—review & editing [Supporting])